oliverschmidt e5665fb968 The recent change turning uip_buf from an array to a pointer seems a quite heavyweight way to have the uip_buf alligned to a 32-bit boundary. Contiki is - still - supposed to be portable across a wide range of toolchains - therefore it i.e. intentionally doesn't use C99 features. From that perspective it seems questionable if all toolchains are able to optimize away the overhead imposed by an "unnecessary" indirection to access uip_buf - and uIP consists mostly of accesses to uip_buf. Even if the future is supposed to bring support for multiple uip_bufs which would turn the uip_buf pointer from a mere workaround to an neceesary element on can still presume that there will be targets with resources limited in a way prohibiting the use of several uip_bufs. So at least for those targets will still be a usecase for a "plain", not indirected uip_buf. ... Therefore it seems reasonable to allow to stay with the "plain" uip_buf at compile-time.		2010-05-08 07:23:55 +00:00
..
cfs
ctk
dev
lib
loader
net	The recent change turning uip_buf from an array to a pointer seems a quite heavyweight way to have the uip_buf alligned to a 32-bit boundary. Contiki is - still - supposed to be portable across a wide range of toolchains - therefore it i.e. intentionally doesn't use C99 features. From that perspective it seems questionable if all toolchains are able to optimize away the overhead imposed by an "unnecessary" indirection to access uip_buf - and uIP consists mostly of accesses to uip_buf. Even if the future is supposed to bring support for multiple uip_bufs which would turn the uip_buf pointer from a mere workaround to an neceesary element on can still presume that there will be targets with resources limited in a way prohibiting the use of several uip_bufs. So at least for those targets will still be a usecase for a "plain", not indirected uip_buf.	2010-05-08 07:23:55 +00:00
sys
contiki-lib.h
contiki-net.h
contiki-version.h
contiki.h